Blogging about your crummy day? Writing a song about unrequited love? Your words may help researchers monitor society's mood swings.
It's not so hard to figure out if one particular person--or a small group of people--is happy or sad. Psychologists hand out questionnaires or conduct interviews. But when it comes to gauging the happiness of an entire society, things aren't so simple. Applied mathematicians Peter Dodds and Christopher Danforth of the University of Vermont in Burlington wondered if the Internet could help. The duo focused on two popular outlets for personal expression: blog posts and song lyrics. Besides providing large slews of data, the researchers believed that people are more honest in personal writings than during formal psychological tests. The scientists analyzed sentences from 2.4 million blogs, which were collected by a Web site called www.wefeelfine.org. The site searches blogs across the world for versions of the phrase "I feel" and then records the entire sentence. For songs, Dodds and Danforth downloaded more than 230,000 lyrics from www.hotlyrics.net, a searchable online database of song lyrics submitted by volunteers.
With the aid of their own computers, the researchers scanned the texts for more than 1000 emotionally charged words that a 1999 psychology study had ranked on a scale from 1 (miserable) to 9 (ecstatic). "Triumphant" and "love" topped the list with average scores greater than 8.7, whereas "disgusted" was one of the lowest at 2.45. The researchers then calculated an average happiness score for each text based on the words' scores and frequencies.
The good news is that blogosphere happiness has increased by about 4% since 2005, the researchers report in an upcoming issue of the Journal of Happiness Studies. When they examined day-to-day fluctuations, they found that certain days stood out: Christmas and Valentine's Day were annual elation peaks at about 6.0, whereas the 11 September anniversaries were gloomy valleys (about 5.7). The happiest day since 2005 was 4 November 2008, the U.S. presidential election, with a score of 6.3. The day's score jumped as bloggers wrote the word "proud" more often and used "pain" and "guilty" less. More recently, Michael Jackson's death triggered a 3-day sadness trough at 5.8.
The researchers also mined the blogs for demographic data. Dodds and Danforth found that men and women had similar average happiness scores (5.89 and 5.91, respectively), but women wrote words at the emotional extremes more often. Teenagers, not surprisingly, were the grumpiest (about 5.5 for 13- and 14-year-olds) and typed words such as "sick," "hate," and "stupid" more often than bloggers aged 45 to 60, the happiest writers.
When the researchers turned to song lyrics, they discovered a darker trend: Popular music has become less happy since 1960--a 10% drop. Most of the decreases happened between 1961 (6.7) and 1980 (6.2). Fewer singers are crooning about "love" now and more are shouting about "hate" and "pain," the scientists report. When they analyzed musical genres, the researchers found that genres haven't changed much over time, but new, less happy genres, such as punk and heavy metal, have become popular and brought lyric scores down. "All-4-One [an R&B group] still uses the word 'love' a lot in their songs," Danforth says. "But [hard rock group] Metallica doesn't use it much at all."
The Vermont scientists are now studying Twitter feeds to watch how emotions change within social networks in real time. James Fowler, a social scientist at the University of California (UC), San Diego, says Dodd and Danforth's methods will enable scientists "to take the pulse of the whole world, assessing the mood of human society." Psychologist Sonja Lyubomirsky of UC Riverside thinks that the data could be used to make interesting predictions. "Do musicians with more positive lyrics sell more CDs?" she asks. "Do [people in] countries with more positive bloggers live longer?"